Research map / Reinforcement learning / Policy gradient and actor-critic
Uncorrected Least-Squares Temporal Difference with Lambda-Return
Takayuki Osogami
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use Boyan's LSTD(lambda) with optimal lambda and alpha instead of Uncorrected LSTD(lambda) with its optimal choices | LSTD(lambda) on random MRPs · small, large, deterministic MRPs with non-binary features | MSE: up to 2 % | Mixed |
| Use Uncorrected LSTD(lambda) rather than Boyan's across the range of lambda and alpha | LSTD(lambda) · random MRPs | MSE stability across hyperparameters | Improved |
| Mix Uncorrected and Boyan's LSTD(lambda) with weight lambda | Mixed LSTD(lambda) · random MRPs | MSE | Mixed |
| Skip the bootstrap correction of Monte Carlo return (Uncorrected instead of Boyan's) in a stateless MRP | LSTD(lambda) · stateless MRP, Proposition 1 | estimator variance | Improved |
| Skip the bootstrap correction of Monte Carlo return in a stateless MRP | LSTD(lambda) · stateless MRP, Proposition 1 | estimator bias | Hurt |
Closest papers by meaning
- The surprising efficiency of temporal difference learning for rare event prediction (NeurIPS 2024 · similarity 0.86)
- Finite Sample Analysis of Linear Temporal Difference Learning with Arbitrary Features (NeurIPS 2025 · similarity 0.85)
- On Double Descent in Reinforcement Learning with LSTD and Random Features (ICLR 2024 · similarity 0.85)
- Almost Sure Convergence of Linear Temporal Difference Learning with Arbitrary Features (NeurIPS 2026 · similarity 0.84)
- Discerning Temporal Difference Learning (AAAI 2024 · similarity 0.84)
- Statistical Efficiency of Distributional Temporal Difference Learning (NeurIPS 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
